Prevalence of common enteric viruses in municipal wastewater treatment plants and their health risks arising from wastewater reuse

Wastewater reuse has been widely practiced for decades worldwide due to the shortage of water resources. Globally, wastewater is widely used in agriculture irrigation, landscape irrigation, and industrial purposes (Wang et al. 2017). It has been reported that up to 18 million hectares of land are irrigated with untreated wastewater (Verbyla et al. 2016), having serious human health implications for both farmers and consumers. In China, given the serious water scarcity and water pollution, water reuse is recognized as an integral part of the water and wastewater management scheme. Thus, the Chinese Government has launched nationwide efforts to optimize the use of reclaimed water (Wang et al. 2017; Dou & Zhu 2018). The quantity of annual recycled and reused wastewater in China reached 3.5 × 10⁹ m³ in 2013 (Wang et al. 2017). However, the health risks induced by human pathogens, especially human viruses, during wastewater reuse have been neglected, and few technical standards governing reclaimed water use have been developed thus far.

Municipal wastewater treatment plants (WWTPs) are important hubs of urban water cycling that receive wastewater from drainage systems. Thus, WWTPs have been considered a great reservoir for human pathogens, including bacteria, protozoa and viruses (Lu et al. 2015; Huang et al. 2018; Farkas et al. 2020). Over 100 types of human viruses have been detected in raw wastewater (Qiu et al. 2015), a large proportion of which are enteric viruses excreted with human stool and urine (Zhong et al. 2007; Aoki et al. 2010). Enteric viruses, including noroviruses (Nov), adenovirus (Adv), astrovirus, enteroviruses (EV), and rotavirus (RV), are primarily transmitted via the fecal–oral route, by either person-to-person contact or ingestion of contaminated food or water (Gibson 2014; Castro-Mayorga et al. 2016). These enteric viruses, most of which are associated with acute gastroenteritis (Rodriguez-Lazaro et al. 2012), present a serious threat to public health for their presumed low infection dose; prolonged (3–4 weeks) and asymptomatic periods of shedding; and excellent environmental stability (Gibson 2014; Kotwal & Cannon 2014). The occurrence and abundance of enteric viruses in wastewater have been monitored in many countries including Japan (Katayama et al. 2008), the UK (Campos et al. 2016), and the USA (Schmitz et al. 2016). However, while WWTPs are effective in reducing viral abundance (Kitajima et al. 2014; Schmitz et al. 2016), information is limited regarding the health risks associated with the reuse of this treated wastewater.

Owing to the substantial hazards of enteric viruses within recycled wastewater, their subsequent spread in the environment should be controlled as much as possible. The abundance of enteric viruses in raw wastewater is reduced in WWTPs via secondary and tertiary treatment processes (Montazeri et al. 2015). However, how much reduction of virus is needed for safe recycled water is still unclear (Gerba et al. 2017). In the USA, California has the most comprehensive water reuse regulations, requiring a 5-log reduction in viruses after the postsecondary treatment prior to unrestricted use of recycled water in irrigation and a 12-log reduction prior to potable reuse via groundwater recharge (Chaudhry et al. 2015). In the case of recycled water used for irrigation, a 6- to 7-log reduction of viruses after the treatment has been suggested by the WHO (Sano et al. 2016). In China, there are no regulations limiting the abundance of enteric viruses, while tolerant concentrations of fecal coliforms (FCs) are specified (Chinese Standard 2010). Recently, quantitative microbial risk assessment (QMRA) has been developed and applied to reasonably evaluate virus-related health risks arising from water reuse (Chhipi-Shrestha et al. 2017; Simhon et al. 2020). QMRA estimates the pathogen dose to which the consumer is exposed, calculates the probability of infection or disease, and finally provides the value of disability-adjusted life years (DALYs) caused by viral infection (Owens et al. 2020). Owing to the common consumption of raw vegetables in many regions, a great number of studies have only considered the health risks caused by exposure to viruses resulting from the agricultural application of wastewater effluent (Moazeni et al. 2017). However, little attention has been paid to public green space irrigation (Chhipi-Shrestha et al. 2017), which is one of the most important wastewater reuse strategies in many cities of China.

This study quantified six types of common enteric viruses in the raw and treated wastewater from three full-scale WWTPs in Nanjing, China, so as to determine the effectiveness of the different treatment processes in reducing viral abundances. Moreover, QMRA was used to estimate the human health risk of two viral types (Adv and norovirus GII (Nov GII)) in the treated wastewater to be recycled for public green space irrigation and crop irrigation. This study deepens our understanding of the prevalence of enteric viruses in WWTPs and will help policymakers to develop science-based regulations for the reuse of treated wastewater.

Wastewater samples were collected from three WWTPs in Nanjing, Jiangsu Province, China, with each WWTP sampled three times from September to December in 2019. Supplementary Material, Table S1 shows the characteristics of the WWTPs. The three plants, which all received domestic wastewater, employed different secondary treatment processes, specifically an oxidation ditch, A²/O and UNITANK. Plant A utilized ultraviolet radiation as a tertiary treatment, while Plants B and C employed chlorination disinfection via sodium hypochlorite and liquid chlorine, respectively. Wastewater samples collected from the WWTPs included the raw wastewater as well as the treated wastewater following both the secondary and tertiary treatment processes.

The raw wastewater (∼5 L/sample) was collected using plastic containers, delivered to the laboratory within 6 h, and then immediately processed for virus concentration upon arrival. The treated wastewater was collected, and the viral fraction from the treated wastewater was concentrated on site using a NanoCeram Virus Sampler (Argonide Corporation) at a rate of approximately 10 L/min (Karim et al. 2009). Water was collected until the membrane fouled beyond the point that the flow rate sharply decreased. Supplementary Material, Table S2 shows the volume of the collected effluent samples. Filters were immediately placed on ice and transported to the laboratory within 6 h.

The enrichment of virus-like particles (VLPs) from the raw wastewater was performed via the electronegative filter method (Schmitz et al. 2016), with slight modification. Turbid samples were first filtered with a Millipore filter membrane (0.45 μm pore size) to remove bacteria, protozoa, and some large particles. Next, 2.5 M MgCl₂ was added to 2 L of filtered samples (final concentration 25 mM MgCl₂). Samples were subsequently passed through an electronegative filter (0.45 μm pore size; catalog no. HAWP-047-00; Millipore, Billerica, MA, USA) attached to a glass filter holder. About 200 mL of 0.5 mM H₂SO₄ (pH 3.0) was passed through the filter to remove magnesium ions, while VLPs were directly absorbed to the filter. Finally, 10 mL of 1.0 mM NaOH was used to elute the VLPs into a tube containing 50 μL of 100 mM H₂SO₄ and 100 μL of 100× Tris–EDTA buffer (pH 8.0) for neutralization. VLP concentrates were stored at −80 °C until further analysis.

The elution of VLPs from the NanoCeram Virus Sampler filter was conducted with 3% w/v beef extract buffer containing 0.05 M glycine (pH 9.5) (Karim et al. 2009). Filters were submerged in 250 mL of elution buffer for 1 min. Next, the inlet and outlet tubes of the filter were connected to peristaltic pump to aid elution for 10 min. A second elution was conducted using the same method as the first elution, except that the submerging time was extended to 15 min. A total of 500 mL of eluate was further concentrated via overnight polyethylene glycol 8,000 precipitation with the final concentration of 120 g/L. After centrifugation for 30 min at 10,000 × g and 4 °C, the supernatant was discarded and the pellets were resuspended in 10 mL of 0.15 M phosphate buffer solution.

Nucleic acids were extracted with the, commercial kit, Allprep^® PowerViral^® DNA/RNA Kit (Qiagen). DNA was directly used for quantitative real-time PCR (qPCR), while RNA was first reverse transcribed to cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) with random primers prior to quantification. The qPCR was performed on the QuantStudio 3 Real-time System (Applied Biosystems) using DNA or RNA samples as templates to detect DNA viruses (human Adv, polyomavirus BK (BK), and polyomavirus JC (JC)) and RNA viruses (human enterovirus (EV), norovirus GI (Nov GI) and Nov GII). The detailed methods (qPCR mixtures, thermal cycling conditions, and quantification methods) are provided in the Supplementary Material (Text S1, and Tables S3 and S4) and described in a previous report (Schmitz et al. 2016). Besides, the recovery efficiency of VLPs during the enrichment process was also taken into consideration for the absolute quantification of viruses (Supplementary Material, Text S1).

The health risk of common enteric viruses was assessed by using the QMRA. This method included four steps: hazard identification, exposure assessment, dose–response assessment, and risk characterization.

Due to limited dose–response assessment data, only Adv and Nov GII were taken into consideration for further health risk assessment. Adv, a double-stranded DNA virus, is widespread in nature and can cause a wide range of infections with a spectrum of clinical manifestations, including in the gastrointestinal tract, respiratory tract, urinary tract, and eyes (Schijven et al. 2019). Nov GII, a single-stranded RNA virus belonging to the Caliciviridae family, is a very important cause of gastroenteritis around the world (Gonzales-Gustavson et al. 2019).

Analysis of variance (ANOVA) was applied to determine whether viral abundances and their reductions were significantly different among the tested samples, and Tukey's test was used for post-hoc analysis. A Monte Carlo simulation of 200,000 iterations was used for the QMRA. Spearman's rank correlation coefficient tests were performed to determine the associations between input parameters and output values of DALYs to describe the sensitivity of input parameters in the QMRA. By changing the input of the LRVs by steps of 0.01 in the QMRA models, we obtained a series of mean DALYs and then used linear regression to build the correlation between the LRV and the mean DALY. All statistical analyses were performed on the R platform (Team RC 2019).

The qPCR was used to determine the abundance of six common enteric viruses (three DNA viruses and three RNA viruses). All six viruses were identified in the raw wastewater, secondary-treated effluent, and tertiary-treated effluent throughout the sampling period in all three Chinese WWTPs (Table 2). Among the DNA viruses, BK exhibited the highest abundance, ranging from 5.38 ± 0.39 to 5.86 ± 0.10 log₁₀ copies/L in the raw wastewater. The mean concentrations of Adv and JC were both >4 log₁₀ copies/L in the raw wastewater from the three WWTPs. In the effluent wastewater, the abundances of Adv and JC were <3 log₁₀ copies/L with the exception of JC in Plant A, while the mean concentrations of BK were still >3 log₁₀ copies/L in Plants A and B. Among the RNA viruses, Nov GII, with a mean concentration of >5 log₁₀ copies/L, was more prevalent than Nov GI with a concentration of about 4 log₁₀ copies/L in the raw wastewater (P < 0.05). However, their concentrations were similar in the effluent wastewater. EV had mean concentrations of >4 log₁₀ copies/L in the raw wastewater from all three WWTPs, and its concentration decreased to 2.86 ± 0.34, 2.45 ± 0.24, and 2.47 ± 0.22 log₁₀ copies/L in the effluent wastewater from Plants A, B, and C, respectively.

To estimate the effect of sampling time and location on the concentrations of the six virus types, ANOVA following two-way factors was performed to obtain the P-values (Supplementary Material, Table S6). No significant differences were observed for the concentrations of EV among the different sampling time points, whereas the concentrations of other viruses showed significant differences among the three WWTPs and the three sampling time points (each P < 0.05). Similarly, the viruses had significantly different concentrations among the three WWTPs and the three sampling times for the effluent wastewater (each P < 0.05), with the exception JC (Supplementary Material, Table S6).

All six viruses decreased after the secondary and tertiary treatments (Figure 1), and the concentrations of FCs showed the similar changes (Supplementary Material, Figure S1). LRVs of the six virus types mostly ranged from 1 to 2 after the secondary treatment. After the tertiary treatment, most viruses exhibited an LRV >2, except Nov GI, which might be due to its relatively lower concentration in raw wastewater. The ANOVA and Tukey's test indicate that the three secondary treatment processes (oxidation ditch, A²/O, and UNITANK) resulted in different reductions of all six virus types, except EV (P < 0.05) (Figure 1). For example, the LRV of BK in Plant C, which used the UNITANK process, reached 3.13 ± 0.32, while its LRV in Plant A, which used an oxidation ditch, was 2.01 ± 0.27. For Adv, BK, JC, and Nov GII, the WWTP with the oxidation ditch showed higher LRVs than the WWTP with A²/O. Similar differences were also observed after the tertiary treatment using three different disinfectants (P < 0.05). Generally, higher LRVs for the viruses were achieved in Plants B and C, which adopted chlorination disinfection, than Plant A adopting ultraviolet radiation. ANOVA also revealed that the six viruses showed significantly different LRVs under the same treatment process (P>0.05, Supplementary Material, Table S7).

Estimated concentrations of enteric viruses in raw wastewater and reduction rates related to the treatment process were obtained with fitted parameters (Supplementary Material, Table S8) and used as inputs for the QMRA models. When the effluent wastewater of the three WWTPs was reclaimed for crop irrigation, the average daily exposure doses were 2.7 × 10⁻⁴–6.1 × 10⁻³ copies and 9.3 × 10⁻³–9.1 × 10⁻² copies for Adv and Nov GII, respectively (Supplementary Material, Figure S2). In contrast, if the wastewater was recycled for public irrigation, the average daily dose of the two viruses was much higher, reaching 1.8 × 10⁻³–4.2 × 10⁻² copies for Adv and 6.4 × 10⁻²–6.3 × 10⁻¹ copies for Nov GII.

Considering exposure frequencies for different reuse scenarios and the morbidity of different viruses, we further obtained the yearly probability of disease for infection by Adv and Nov GII (Supplementary Material, Figure S3). The yearly probability of disease was much lower for crop irrigation than that for public irrigation. Moreover, disease via Nov GII infection showed a higher yearly probability than that infected by Adv. We further considered the disease burden and the proportion of the population susceptible to the disease to obtain the DALYs, in which case we could compare the health risk of the two viruses under a uniform criterion (Figure 2). The DALYs of both viruses were much higher than the WHO-recommended threshold of 10⁻⁶, irrespective of whether the treated wastewater was reused for public or crop irrigation. For public irrigation, the mean DALY values were 3.3 × 10⁻⁴–3.4 × 10⁻² for Adv and 4.0 × 10⁻⁴–5.4 × 10⁻⁴ for Nov GII. For crop irrigation, the DALY values decreased to 2.9 × 10⁻⁴–6.1 × 10⁻³ for Adv and 7.8 × 10⁻⁵–3.9 × 10⁻⁴ for Nov GII. Although Adv had relatively lower values for the daily exposure dose and yearly probability of disease, it exhibited significantly higher DALYs than Nov GII due to the greater disease burden of Adv.

Similarly, significant differences of DALYs were observed among the three WWTPs in this study (P < 0.05). The health risk of Adv arising from wastewater reuse was lowest for Plant C and highest for Plant B. In comparison, the health risk of Nov GII was lowest for Plant B and highest for Plant A. Furthermore, sensitivity analysis demonstrated that the disease burden and proportion of the population susceptible to the disease were the most sensitive input parameters affecting the results of DALYs in the three WWTPs (Figure 3), regardless of virus type and reuse scenario. With the aid of models, we determined the regression equation between the DALYs and LRV for the treatment processes in the WWTPs (Supplementary Material, Table S9 and Figure S4) in order to obtain the required LRV in the WWTPs to meet the requirement of 10⁻⁶ DALYs, as proposed by the WHO (Victor et al. 2008). If the treated wastewater is reused for crop irrigation, the LRVs for Adv and Nov GII should reach 5.51–5.87 and 4.60–4.97, respectively (Table 3), while higher LRVs should be reached for both Adv and Nov GII (Table 3) to meet the reuse requirements for public irrigation.

The quantification of enteric viruses in the wastewater was the first and the most fundamental step in the assessment of health risks with QMRA. Considering the relatively low levels of virus in the wastewater, especially in the effluent after the tertiary treatment process, enrichment is essential before the quantification. In this study, we adapted the method with an electronegative filter for the enrichment of VLPs in the raw sewage (Katayama et al. 2008; Schmitz et al. 2016). However, this method could only handle with the samples with small volumes; thus, it could not be applied to enrich VLPs in the effluent after secondary and tertiary treatments because of extremely low virus levels. We turned to another enrichment method with the NanoCeram Virus Sampler filter which could concentrate a larger volume of water to meet the detection limit though much more costly and time-consuming (Pang et al. 2012; Cashdollar & Wymer 2013). Neglection of the recovery efficiency during enrichment would cause the relatively low abundance of viruses and thus led to underestimate the health risk. In the present study, we used recovery efficiency obtained from spiking experiments to adjust the absolute quantification of viruses to get comparatively accurate results. Besides, the recovery efficiency of viruses during nucleic acid extraction, reverse transcription, and qPCR (extraction–RT–qPCR efficiency) was another factor that might underestimate virus abundances. However, the recovery efficiency during the process was very close to 100% according to a previous study, in which murine norovirus was used as sample process control to determine the efficiency (Schmitz et al. 2016). Thus, in the present study, the quantification of viruses was not adjusted with extraction–RT–qPCR efficiency.

All six viruses are commonly detected in WWTPs around the world. In the present study, polyomavirus BK, which is known to cause hemorrhagic cystitis in immunocompromised patients that receive bone marrow transplantation (Ahasan et al. 2019), showed a high prevalence in the raw wastewater. Urine is considered the main excretion mechanism by infected patients (Zhong et al. 2007). Similar BK concentrations (5–6 log₁₀ copies/L) were also observed in wastewater from two WWTPs in Japan (Kitajima et al. 2014). The abundance of Adv in the untreated wastewater was also similar to the data reported in Arizona (USA) (Schmitz et al. 2016) and London (UK) (Purnell et al. 2016), but was lower than that in California (Chaudhry et al. 2015).

Additionally, we found that Nov GII was the dominant RNA virus during the autumn and winter in the raw wastewater from the WWTPs and showed a significantly higher abundance than Nov GI. Different prevalences of Nov GI and Nov GII have also been found in some WWTPs located in England (Campos et al. 2016), while WWTPs in the USA (Schmitz et al. 2016) and Japan (Katayama et al. 2008) did not show any difference between the noroviruses. In China, Nov GII, which can cause acute gastroenteritis, is the most predominant genotype infecting humans (Ao et al. 2017). It has been reported that 556 norovirus outbreaks occurred between October 2016 and September 2018, with 81.2% of all norovirus outbreaks typed as GII.2[P16] (Jin et al. 2020). Between 2012 and 2016, more than half of the outbreaks were reported to occur during November and December (Jin et al. 2020). High viral loads can usually be detected in the stool samples of infected patients, and these viruses can subsequently enter WWTPs by the urban drainage system (Strubbia et al. 2019). Although the duration of symptoms induced by norovirus infection usually ranges from 1 to 3 days, the average period in which norovirus is excreted in the stool is 14.3 days (Aoki et al. 2010).

This study indicates that the secondary treatment plays an important role in the reduction of viruses from all the three WWTPs, which agrees with previous studies (Francy et al. 2012; Campos et al. 2016). However, the different secondary treatments, namely oxidation ditch, A²/O and UNITANK, were found to have significantly different LRVs for the viruses. Many factors associated with the secondary treatment may affect the removal of viruses, including adsorption to suspended particles, biological predation within the microbial community, and deactivation by solar light or enzymes (Katayama et al. 2008; Chaudhry et al. 2015). Moreover, the mixed liquid suspended solids and hydraulic retention time of the system can exert a great influence on the reduction of virus (Katayama et al. 2008; Chaudhry et al. 2015). Even so, due to the lack of information regarding the removal mechanisms of viruses, it is difficult to determine the contribution of different external factors to the reduction efficiency. Thus, further optimization of operation parameters to improve virus reduction by the secondary treatment process remains challenging.

Disinfection is commonly adopted as a tertiary treatment process in municipal WWTPs to further prevent the spread of human pathogens (Gerba & Pepper 2019). In this study, we found that ultraviolet disinfection provided limited reduction of most tested viruses when compared with chlorination disinfection, which is consistent with a previous study conducted in Canada (Qiu et al. 2015). Although ultraviolet irradiation technology has been widely used for wastewater disinfection due to its safety and economic benefits, viral reduction rates highly depend on the quality of the wastewater, especially with respect to turbidity (Hijnen et al. 2006; Gray et al. 2014). Compared with ultraviolet disinfection, which mainly targets nucleic acids and renders the genome nonreplicable, chlorination disinfection can simultaneously damage genomes and proteins, and inhibit host-cell recognition/binding (Wigginton et al. 2012). In addition, previous studies have also shown that ultraviolet disinfection is more effective in reducing bacterial pathogens than virus, especially Adv (Gray et al. 2014; Song et al. 2016). Thus, a greater ultraviolet irradiation dose is needed to meet the demand of virus reduction.

Insufficient reduction of viruses is an important cause of high health risks associated with wastewater reuse. Though the utilization of recycled water is encouraged in China, little attention has been paid to the health risks induced by virus contamination during the recycling. A standard for landscape irrigation using recycled water (GB/T 25499-2010) was released in 2010 (Chinese Standard 2010). However, the only guideline concerning pathogens is FCs, with a threshold of 200 CFU/L for unrestricted access green spaces, such as those of public parks, campuses or communities. It is a more restricted standard when compared with the Discharge Standard of Pollutants for Municipal WWTPs (Chinese Standard 2002). In the present study, FCs were not detected in the final effluent of Plants B and C, while the mean FC concentration in Plant A was 291 CFU/L, which is slightly higher than the guideline for irrigation (Supplementary Material, Figure S1). Notwithstanding, the effluent from the three WWTPs did not satisfy the WHO-recommended threshold from the point of view of health risks. Indeed, the LRVs for Adv and Nov GII should be higher than 6 and 5, respectively. Thus, advanced treatment processes, such as membrane filtration, should be supplemented in WWTPs for recycling water in order to obtain further viral reduction. Direct ultrafiltration using membranes with nominal molecular weight cutoffs of 1–100 kDa can effectively remove viruses through size exclusion. For example, an LRV of >4 was achieved using a membrane with a nominal molecular weight cutoff of 1 kDa (Shirasaki et al. 2017). One study conducted in Canada also found that after secondary treatment processes, membrane ultrafiltration with a nominal pore size of 0.04 μm could achieve a log reduction of 4.85 ± 0.59 for Adv and 4.55 ± 0.66 for Nov (Qiu et al. 2015).

Different DALYs of the viruses associated with public irrigation and crop irrigation indicate that the proper reuse of recycled water is important for controlling the health risk. When the recycled water was reused for public irrigation, higher DALYs were observed when compared with crop irrigation. Although people are exposed to a larger volume of water, the number of viruses attached to lettuce might be greatly reduced in the field and may undergo additional slight reduction because of storage and washing (Yates et al. 1987; Beuchat 2002; Li et al. 2015). In contrast, people are more directly exposed to the viruses via ingestion and plant contact when the recycled water is used for irrigation of public green spaces (Page et al. 2014; Simhon et al. 2020). Furthermore, consumption of raw leafy vegetables is uncommon in Chinese dietary habits. Thus, the DALYs caused by crop irrigation may be overestimated. By comparison, a common reuse scheme implemented in China is to irrigate public green spaces and recharge into rivers and groundwater (Yi et al. 2011).

Besides, Adv showed to have higher DALYs than Nov II in the sampling period for the different reuse scenarios. As the outbreaks of Nov II usually occur during September and December (Jin et al. 2020), the abundance and the induced health risk of Nov II might reach the peak in this period. However, Adv does not share the same prevalence periodicity with Nov II since the outbreaks of Adv can take place in any season and there is no time specificity for its infections (James et al. 2007; Biggs et al. 2018; Killerby et al. 2019). Thus, temporal variation of the abundance of viruses in WWTPs should be deeply understood for effective control of the induced health risk.

We conducted a detailed evaluation on the health risks of viruses in treated wastewater for recycling based on a QMRA framework; however, this study still had some limitations, especially regarding dose–response models and input parameters concerning disease burden. The variation of dose–response models and their parameters were not considered, which might cause epistemic uncertainty of the DALYs (Donald et al. 2011). Thus, we adopted updated models of Adv and Nov GII dose–responses proposed by Teunis et al. (2008, 2016), which have been used in many published studies (Poma et al. 2019; Schijven et al. 2019). Additionally, sensitivity analysis showed that the disease burden and proportion of the population susceptible to the disease had the greatest impact on the DALY results. Owing to the lack of related studies in China, the disease burden distributions were sourced from studies conducted in Canada and Australia (Gibney et al. 2014; Chhipi-Shrestha et al. 2017). Thus, further studies on the input parameters for exposure assessment, dose–response models, and risk characterizations should be carried out to increase the accuracy and reduce epistemic uncertainty of QMRA results.

This study investigated the abundance of common enteric viruses in raw wastewater and the effectiveness of different secondary and tertiary treatment processes in reducing viral abundances in three full-scale WWTPs in the same city of China. Polyomavirus BK and Nov GII exhibited the highest abundances among the detected DNA and RNA viruses, respectively, with concentrations >5 log₁₀ copies/L in the raw wastewater. Viruses in the raw wastewater were mainly removed by the secondary treatment processes in the three WWTPs, and the log reduction values ranged from 1 to 2. Tertiary treatment processes including chlorination disinfection and ultraviolet disinfection could provide additional reduction of viruses. When the recycled water was used for public or crop irrigation in actual scenarios, the estimated DALYs were above the WHO-recommended threshold of 10⁻⁶ for both Adv and Nov GII. Thus, more effective treatment technologies must be implemented to remove viruses to meet the health requirements proposed by the WHO.

This work was financially supported by the National Natural Science Foundation of China (52025102), the Key Projects of Chinese Academy of Engineering (2020-ZD-15), the Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX18_0034), the Excellent Research Program of Nanjing University (ZYJH005), and the Program B for Outstanding PhD Candidate of Nanjing University.

All relevant data are included in the paper or its Supplementary Information.